img_uuid, then poll or use callbacks. GoEnhance provides independent access and is not Alibaba's official API.Qwen-Image 3.0 Pro API

One Endpoint for Generation and Editing
Generate from a Prompt
image_list for pure text-to-image generation. Choose a supported ratio and 1K or 2K output, or let ratio: auto select the composition.
Edit with Up to Three References
image_list. Their order controls how the prompt refers to each source; every reference must be under 10MB.
Qwen-Image 3.0 Pro API Parameters
Required Model
model to qwen-image-3-pro.Prompt
Aspect Ratio
auto or one of 13 documented ratios from 1:1 through 8:1.Output Tier
image_size to 1K or 2K; the default is 2K.Reference Images
image_list, each under 10MB.Prompt Controls
prompt_extend_mode. The agent mode is text-to-image only.Qwen-Image 3.0 Pro API Pricing
1K Output
2K Output
ratio: auto are billed at this tier.Reference Surcharge
1:8 and 8:1 are limited to—and billed at—the 1K tier.Integrate Qwen-Image 3.0 Pro in Three Steps
1. Build and Validate
2. Submit and Persist
model: qwen-image-3-pro and save the returned img_uuid.3. Handle Completion
GET /api/v1/jobs/detail or accept HTTPS callback events until the task succeeds or fails, then store the output or handle the error.Qwen-Image 3.0 Pro API Use Cases
Product Visuals
Ad Variations
Design Prototypes
Reference-Guided Restyling
Localized Campaigns
Content Pipelines
Qwen-Image 3.0 Pro API FAQ
How much does the Qwen-Image 3.0 Pro API cost?
GoEnhance documents 2.43 tokens ($0.049) per 1K output and 4.85 tokens ($0.097) per 2K output, plus 0.19 tokens ($0.0039) for each reference image.
Is this Alibaba's official API?
No. Qwen is associated with Alibaba, while this page describes independent API access provided by GoEnhance. It does not imply partnership or endorsement. Developers who require an Alibaba-operated API should consult Alibaba Cloud documentation.
Can the endpoint generate and edit images?
Yes. Omit image_list for text-to-image generation, or include one to three reference-image URLs for editing. The endpoint returns one generated image.
Which ratios are supported?
The documented values are auto, 1:1, 16:9, 9:16, 4:3, 3:4, 2:3, 3:2, 1:2, 2:1, 1:4, 4:1, 1:8, and 8:1.
How does prompt extension work?
The default direct mode uses the prompt directly. agent rewrites more thoroughly but is only valid for text-to-image requests without image_list.
Does the API return the image immediately?
No. A successful submission returns an img_uuid. Poll the job-detail endpoint or use callback events to retrieve the terminal result.
What happens if my callback is unavailable?
If your endpoint does not return HTTP 200, GoEnhance documents up to three callback retries with a three-second timeout per attempt. Polling remains a practical recovery path.
